Improved Stokes–Einstein based models for diffusivities in supercritical CO2

Abstract

The large interest and applications of supercritical carbon dioxide (SC-CO2) require the existence of transport properties for simulation and/or design. In this work, pure predictive models are proposed for the accurate estimation of binary diffusivities at infinite dilution in SC-CO2. They are three simple and straightforward expressions grafted on the Wilke–Chang, Scheibel and Lusis–Ratcliff equations, whose deviations to the Stokes–Einstein behavior in the supercritical domain are corrected by introducing two universal constants. Such modifications decrease the average errors from [11.70–23.16]% to [8.26–8.51]% for a large database of 150 systems and 4484 data points over wide ranges of temperature and pressure. The dispersion of the computed errors is also significantly lower, which highlights the reliability of our improved models to accurately predict tracer diffusivities.